See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/225243670

The Polymorphism of Glycine. Thermochemical and structural aspects

Article  in  Journal of Thermal Analysis and Calorimetry · December 2001

DOI: 10.1023/A:1013179702730

CITATIONS READS

187 529

3 authors:

German L Perlovich Lars Kristian Hansen

Institute of Solution Chemistry of RAS UiT The Arctic University of Norway
248 PUBLICATIONS   3,748 CITATIONS    65 PUBLICATIONS   1,690 CITATIONS   

SEE PROFILE SEE PROFILE

Annette Bauer-Brandl
University of Southern Denmark
134 PUBLICATIONS   3,279 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Drug-drug cocrystals of antituberculous 4-aminosalicylic acid: screening, crystal structures, thermochemical and solubility studies View project

Isolation and characterization of bioactive compounds from medicinal plants View project

All content following this page was uploaded by German L Perlovich on 02 January 2015.

The user has requested enhancement of the downloaded file.

 Journal of Thermal Analysis and Calorimetry, Vol. 66 (2001) 699–715

THE POLYMORPHISM OF GLYCINE

Thermochemical and structural aspects

G. L. Perlovich1, 2, L. K. Hansen3 and A. Bauer-Brandl1*

1
University of Tromsø, Institute for Pharmacy, Breivika, N-9037 Tromsø, Norway
2
Institute of Solution Chemistry, Russian Academy of Sciences, Ivanovo, Russia
3
University of Tromsø, Department of Chemistry, Breivika, N-9037 Tromsø, Norway

(Received March 9, 2001; in revised form May 3, 2001)

Abstract
X-ray, DSC and solution calorimetric investigations were carried out for α-, β- and γ-modifications
of glycine. Particular attention was paid to kinetic and thermochemical aspects of γ- → α-phase tran-
sition. The temperature of this phase transition turned out to be sensitive to a) conditions under
which the crystals of the γ-modification were grown, b) tempering of crystals c) form (geometry) of
crystals. Kinetics of this phase transition of single crystals of γ-phase in rhomboedric form can be de-
scribed by the equation for two-dimension nuclei growth, whereas for crystals of triangle geometry
the equation for three dimension growth is valid. On the basis of energy parameters describing
growth of α-form in γ- → α-phase transition, the kind of structure defects, which are responsible for
this phase transition, was estimated. Taking into account the ∆solHm, the absolute values of the lattice
energies of the investigated polymorphs in descending order are follows: γ->α->β-modification. The
obtained results are discussed with respect to the peculiarity of the crystal lattice structures, particu-
larly the network of hydrogen bonds. The β-modification of glycine is monotropically related to the
other forms, whereas γ- and α-polymorphs are enantiotropically-related phases.

Keywords: crystal structure, enthalpies of solution, enthalpy of phase transition, glicine, kinetics
of phase transition, monotropic and enantiotropic phases, phase transition, polymor-
phism, single crystal

Introduction
Three different polymorphic phases of glycine have already been known for a long
time: the first attempts to study α-glycine (α-Gly) by means of X-ray diffraction were
undertaken by Bernal [1] and by Hengstenberg and Lenel [2] independently. How-
ever, accuracy and interpretation of their results left something to be desired. The ex-
plicit crystalline structure was refined by Albrecht and Corey [3], and a precise re-
finement of not only the heavier atoms but also locating of the hydrogen atoms for

* Author for correspondence: E-mail: annetteb@farmasi.uit.no

1418–2874/2001/ $ 5.00 Akadémiai Kiadó, Budapest

© 2001 Akadémiai Kiadó, Budapest Kluwer Academic Publishers, Dordrecht
700 PERLOVICH et al.: THE POLYMORPHISM OF GLYCINE

α-modification was carried out by Marsh [4]. In the seventies, Jönsson and Kvick [5]
investigated deformation electron density of this modification employing X-ray and
neutron diffraction measurements.
β-Glycine (β-Gly) had already been obtained and described by Fischer [6] at the
beginning of the century, whereas X-ray examination was performed quite a long
time later [1, 7]. It can be suspected that this fact was due to the general low stability
of this phase and possibly due to irreversible transformation into α- or γ-Gly forms.
γ-Glycine (γ-Gly) was first revealed and the structure resolved by X-ray diffrac-
tion by Iitaka [8–10]. Moreover, he analysed not only the packing of the molecules in
the crystal lattice, but also the role of the hydrogen bonds in the formation of the
framework and the substructure of the hydrogen bonds. Investigations of γ-glycine at
298 and 83 K were carried out by means of neutron diffraction by Kvick [11] in order
to determine electron density changes of the molecules with temperature.
Numerous studies dealing with thermophysical and thermochemical studies of
glycine are literally known: as examples may be mentioned determination of combus-
tion enthalpies [12–14], constant pressure heat capacity of the solid [15–17], and sub-
limation enthalpies [13, 18–20]. Unfortunately, none of these articles take into ac-
count the crystal lattice of the studied phases. Furthermore, investigations into the
thermodynamics of phase transitions of the considered polymorphs are not practi-
cally existent.
In the present study, thermochemical aspects of the phase transition γ- → α-Gly
are studied in detail. Furthermore, differences of crystal lattice energies of the poly-
morphs are investigated.

Material and methods

Glycine (aminoacetic acid, >99%) was from Sigma (Sigma-Aldrich Sweden AB,
Stockholm, Sweden). For crystallisation of the polymorphs the following solvents
were used: ethanol, extra pure grade (99.6 v/v%, maximum water content 0.4%); ace-
tic acid (analysis grade, 100%), bought from Merck (Merck GmbH, Darmstadt, Ger-
many) and distilled water.
All the polymorphs under examination were grown as described earlier [7, 10].
The obtained crystals were thoroughly dried under vacuum and stored under mois-
ture-free conditions. Details of thermal treatment and preparation of the crystals are
described below.
Thermal properties of different modifications of glycine were measured using a
Perkin Elmer Pyris 1 DSC differential scanning calorimeter (Perkin Elmer Analytical
Instruments, Norwalk, Connecticut, USA). DSC runs were performed in an atmo-
sphere of flowing dry nitrogen gas of high purity (20 ml min–1) using closed standard
aluminium sample pans. DSC was calibrated with indium from Perkin Elmer (P/N
0319-0033). The value of the determined enthalpy of fusion corresponded to
–28.48 J g–1 (reference value –28.45 J g–1). The melting point was 156.5±0.1°C.
Enthalpies of solution (∆solHm) were measured using a Precision Solution Calo-
rimeter in the 2277 Thermal Activity Monitor Thermostat (both from Thermometric

J. Therm. Anal. Cal., 66, 2001

 PERLOVICH et al.: THE POLYMORPHISM OF GLYCINE 701

AB, Järfälla, Sweden). The software SolCal Version 1.2 (Thermometric) was applied
to all calculations. The measuring temperature was 25±10–4°C, volume of the vessel
100 ml, stirrer speed 500 rpm and the mass of investigated compounds approximately
63 mg. The accuracy of mass measurements corresponded to ±10–6 g. The calorimeter
was calibrated using KCl (analysis, grade >99.5%, from Merck) in water in a wide
concentration interval with a number of measurements of more than 10. The obtained
standard value of solution enthalpy was ∆solH0=17 225±50 J mol–1. This is in good
agreement with the value recommended by IUPAC of ∆solH0=17 217±33 J mol–1 [21].
Single-crystal X-ray measurements were carried out using a Nonius CAD-4
diffractometer with graphite-monochromated MoKα radiation (λ=0.71069). Intensity
data were collected at 20°C by means of a ω-2θ scanning procedure up to a 2θ limit of
70°. The crystal structure was resolved using direct methods and refined by means of
a full-matrix least-squares procedure. All programs used in the solution (Sheldrick,
1997a [22]), refinement (Sheldrick, 1997b [23]) and display [24] of the structures are
included in the OSCALL program package [25].

Results and discussion

Crystal structures of different polymorphic modifications of glycine
The crystal structure of a polymorph determines its lattice energy. Therefore, before
discussing experimental results, details of crystal structures of α-, β- and γ-modifica-
tions of glycine are discussed. The parameters of the investigated crystal lattices ob-
tained by other and the present authors (in order to characterise the single crystals un-
der investigation) using various X-ray and neutron diffraction methods are presented
in Tables 1and 2. Final fractional co-ordinates are listed in Table 3.
All data derived for the three modifications in this study are in good agreement
with literature data [1–7, 10, 11].
The dimension of the carboxyl group and the arrangement of oxygen atoms
around the nitrogen atom suggest that the molecules in all three (α-, β- and γ-) modi-
fications form zwitter ions. The configuration states of the molecules in the consid-
ered polymorphic phases differ from one another only by the angle between C–N and
the least-squares plane (a twist around the CI–CII bond): 18.6° in α-; 24.8° in β-; 12.8°
in γ-phases (Fig. 1). This fact is a consequence of the complicated packing of mole-
cules in the crystal lattice and the competition of several kinds of interactions: a) van
der Waals, b) electrostatic (particularly dipole-dipole interaction) and c) framework
hydrogen bonds.
In order to explain experimental results of different methods (IR, X-ray), partic-
ular attention is paid to the structure hydrogen bonds. Therefore, in order to analyse
the number and strengths of the hydrogen bonds located between a reference mole-
cule and its nearest neighbours, the scheme proposed by Iitaka is used [10]. The pro-
jection of molecule bonds of the modifications is presented in Fig. 2, where the direc-
tion along the C–N bond is taken as the polar axis (N–H bond also indicated).

J. Therm. Anal. Cal., 66, 2001

J. Therm. Anal. Cal., 66, 2001

702
Table 1 The lattice parameters of the α-phase of glycine

Marsh [4] Albrech and Corey [3] Bernal [1] Jönsson and Kvick [5]b Legros and Kvick [36] This workc

PERLOVICH et al.: THE POLYMORPHISM OF GLYCINE

a/Å 5.1020(8) 5.10 5.04 5.1054(6) 5.0835(10) 5.107(2)
5.0830(8)b
b/Å 11.9709(17) 11.96 12.1 11.9688(19) 11.820(2) 12.040(2)
11.816(3)b
c/Å 5.4575(15) 5.45 5.41 5.4645(9) 5.4579(9) 5.460(2)
5.4590(10)b
α/deg 90 90 90 90 90 90
β/deg 111.71(2) 111.63 111.63 111.697(11) 111.95(2) 111.82(2)
111.93(2)b
γ/deg 90 90 90 90 90 90
Z 4 4 4 4 4 4
3
V/Å 309.69 309.02 306.69 310.25 304.17 311.67(17)
304.15b
SGa P21/n P21/n P21/n P21/n P21/n P21/n
T/K 298 298 298 298 120 298(2)
a
space group
b
neutron diffraction
c
after DSC measurement; number of reflections measured are 544; number of independent reflections are 489; absorption coefficient is µα =0.142 mm–1;
final R-value is 0.059
 Table 2 The lattice parameters of the β- and γ-phases of glycine

PERLOVICH et al.: THE POLYMORPHISM OF GLYCINE

β-glycine γ-glycine
c b
Bernal [1] Iitak [7] This work Iitaka [10] Kvick [11] Kvick [11]b This workd
a/Å 5.18 5.077(4) 5.0943(13) 7.037 7.046(3) 6.975(2) 7.035(2)
b/Å 6.18 6.267(6) 6.286(5) 7.037 7.046(3) 6.975(2) 7.035(2)
c/Å 5.29 5.379(9) 5.3831(15) 5.483 5.491(2) 5.473(2) 5.481(2)
α/deg 90 90 90 90 90 90 90
β/deg 114.33 113.2 113.21(2) 90 90 90 90
γ/deg 90 90 90 120 120 120 120
Z 2 2 2 3 3 3 3
3
V/Å 154.31 157.36 158.44(13) 235.14 236.1(2) 230.6(2) 234.91(13)
SGa P21 P21 P21 P32 P32 P32 P32
T/K 298 298 298 298 298 83 298(2)
a
space group
J. Therm. Anal. Cal., 66, 2001

b
neutron diffraction
c
number of reflections measured are 374; number of independent reflections are 353; absorption coefficient is µβ-= 0.140 mm–1; final R-value is 0.046
d
number of reflections measured are 746; number of independent reflections are 741; absorption coefficient is µγ-= 0.141 mm–1; final R-value is 0.036

703
J. Therm. Anal. Cal., 66, 2001

704
PERLOVICH et al.: THE POLYMORPHISM OF GLYCINE
Table 3 Final position of atoms in three glycine modifications

α-phase β-phase γ-phase

x y z x y z x y z
N –0.3008(8) 0.4104(3) 1.2588(7) –0.3509(6) 0.0129(5) 1.2623(5) 0.75431(19) 0.78341(19) 0.6222(2)
O1 –0.3042(6) 0.4055(2) 0.7653(5) 0.0914(4) 0.1405(5) 0.9029(4) 0.76732(18) 0.77862(18) 0.1328(2)
O2 0.1472(6) 0.3586(3) 0.8954(6) –0.3780(5) 0.0897(6) 0.7575(4) 0.4570(2) 0.4565(2) 0.1004(2)
C1 –0.0752(8) 0.3750(3) 0.9341(7) –0.1122(6) 0.1267(7) 1.2305(5) 0.6070(2) 0.6080(2) 0.2208(2)
C2 –0.0647(9) 0.3549(3) 1.2130(8) –0.1337(6) 0.1157(5) 0.9391(5) 0.6014(3) 0.5769(2) 0.4967(3)
H1 0.101(9) 0.386(3) 1.332(9) 0.051(9) 0.064(8) 1.376(8) 0.665(7) 0.478(7) 0.547(8)
H2 –0.086(9) 0.271(4) 1.234(9) –0.112(9) 0.261(10) 1.286(9) 0.455(6) 0.522(7) 0.570(10)
H3 –0.492(12) 0.387(4) 1.140(11) –0.313(15) –0.116(19) 1.210(15) 0.886(4) 0.823(4) 0.576(5)
H4 –0.303(11) 0.490(5) 1.229(11) –0.526(10) 0.047(7) 1.159(9) 0.724(4) 0.889(4) 0.589(6)
H5 –0.295(9) 0.398(4) 1.432(10) –0.331(10) 0.039(9) 1.412(10) 0.708(7) 0.744(6) 0.806(8)
 PERLOVICH et al.: THE POLYMORPHISM OF GLYCINE 705

Fig 1 Conformation state of molecule in the crystal lattice for the different glycine modifications

Fig 2 Stereographic projections of NH3–CII fragment of glycine molecule (along N–CII

bond) and nearest oxygen atoms: a – α-, b – β-, c – γ-modifications

The molecules, received by a threefold screw axis of a molecule M, are desig-

nated by T1 and T2 and other molecules which are related to these three by lattice
translations are designated with subscript the translation vector (for example Ma, Tab1 ).
In α- and β-glycine two oxygen atoms (OI c and OIIMa ) are situated approximately in
M

the tetrahedral directions and are participating in those strong hydrogen bonds which
hold the molecules in a sheet. The third hydrogen atom (NH3+-group) is situated on a
W W
line approximately bisecting the angle OIMa –N–OIIM (α-Gly) or OI a b –N–OII b (β-Gly)
and forms a bifurcate hydrogen bond, which accounts for the cohesion of the sheets.
For the γ-phase the situation is different: five oxygen atoms at short distances sur-
round the nitrogen atom, which
1
implies the presence of hydrogen bonds. The three
oxygen atoms OIMc , OIIT and OITab in the γ-modification are arranged approximately in
1

the tetrahedral directions. Therefore, these atoms

2
are taking
2
part in hydrogen bonding
and the rest of the oxygen atoms, namely OITa and OIITab , interact with NH +3 -group by
means of electrostatic attraction.
The respective structures of α- and β-Gly are shown in Figs 3a and 3b accord-
ingly. They may be described as structures consisting of molecular layers lying in
parallel to the (010) planes. The configuration of the sheets in these two polymorphs
is almost the same. In β-Gly, the single sheets are packed together by hydrogen bonds
in the b-direction, whereas in α-Gly, pairs of two single sheets are held together by
hydrogen bonds to form double sheets, which then are packed together by van der
Waals forces.

J. Therm. Anal. Cal., 66, 2001

706 PERLOVICH et al.: THE POLYMORPHISM OF GLYCINE

The framework of hydrogen bonds in the γ-Gly crystal lattice is schematically

shown in Fig. 3c. The dominating feature of the intermolecular packing is the hydro-
gen bonding between amino atoms which links the molecules into helical chains
around the threefold screw axes parallel to the c-axis. These chains are packed to-
gether by lateral hydrogen bonds forming a three-dimensional network of hydrogen
bonds. The electrostatic forces between oppositely charged groups would enhance
the effect of such bonding. Furthermore, the alignment of the hydrogen bonds seems
to be better in this case than with the corresponding hydrogen bonds in the other
forms of glycine. As derived from IR-experiments [7, 26], in α- and β-glycine the lat-
eral hydrogen bonds are stronger than the interlayer hydrogen bonds. It is interesting
to note that the methylene–methylene or methylene–oxygen distances in the γ-phase
are somewhat longer than observed between the double layers of α-Gly, and the

Fig 3 Schematic illustration crystal structure and network of hydrogen bonds for differ-
ent polymorphic forms of glycine

J. Therm. Anal. Cal., 66, 2001

 PERLOVICH et al.: THE POLYMORPHISM OF GLYCINE 707

packing of the molecules within the crystal is less compact compared to the α-form.
Presumably, hydrogen bond and electrostatic energies overcome the decrease in van
der Waals energies.
Theoretical calculation of values of crystal lattice energies is a rather demanding
problem because of the competition of several energetic terms. Moreover, it is diffi-
cult to estimate a bifurcate hydrogen bond by means of mathematical procedures.
In the present study, attempts were made to measure the differences of crystal
lattice energies of the investigated polymorphs by solution calorimetry as a direct ex-
perimental method.

DSC measurements
The γ- → α-phase transition
The first scientist who studied the transition of glycine phases by means of DSC ex-
periments was Iitaka [8, 10]. It should be noted that he emphasised the unusual be-
haviour of the heat effects he measured. First of all, they were widely spread over a
wide temperature range of 165±5°C. Secondly, this phase transition was very sensi-
tive to grinding and other mechanical treatment of the crystals. For example, the
phase transition takes place when the crystals of the γ-phase are powdered. As a con-
sequence of this fact, the phase transition (both the transition temperature and the
heat effect) is sensitive to the function of distribution of structure defects in the crys-
tal and, consequently, to the procedure of preparation of a crystal. Kvick [11] has
shown that the structure of γ-glycine contains mosaic spreads of17.4′′ (83 K) and 208 . ′′
(298 K). This fact is also an evidence for the relation between growth parameters and
perfection of a crystal.
The preparation of single crystals of γ-glycine usually takes place at a tempera-
ture around 1°C. However, experimental temperatures (DSC and solution calorime-
try) exceed this temperature. Hence the density of different kinds of structure defects
does not have the thermodynamical equilibrium value. Taking this into account, an
exceeded number of defects will relax during time and at definite environmental con-
ditions the system (crystal+defects) achieves a stage closer to equilibrium. As far as a
relaxation process is a kinetic process, the relaxation time depends on the temperature
treatment. This can be used for an effective homogenisation of properties in a crystal.
Moreover, dislocations and other defects of the crystal structure promote local disor-
der, even breaking of the network of the hydrogen bonds, particularly those weak bi-
furcate and lateral hydrogen bonds. Therefore, reproducibility of DSC and solution
calorimetry measurements can be widely improved by thermal treatment of the mate-
rial (tempering) prior to the measurements without changing crystal lattice.
In order to confirm this hypothesis the following experiment was carried out: a
perfect γ-glycine single crystal with a cross area of approximately 2 mm2 (which had
three big clear-determinated faces) was carefully divided into several equal pieces.
Two pieces were used to examine the γ- → α-phase transition temperature immedi-
ately upon breaking by means of DSC at a heating rate of 10 K min–1. The remaining
pieces were exposed to thermal treatment at 175°C during 4.5 h under a dry nitrogen

J. Therm. Anal. Cal., 66, 2001

708 PERLOVICH et al.: THE POLYMORPHISM OF GLYCINE

atmosphere, prior to DSC measurement. The temperatures of phase transition were

essentially different: for the untreated material Ttr=189.3°C and for the treated mate-
rial Ttr=200.7°C respectively. Similar dependencies were observed for the crystals of
other forms and geometry. For example, the initial (fresh prepared) rhombus-like
crystals had Ttr=168.4°C, whereas the phase transition temperature of these crystals
after thermal treatment at 159°C for 5 h was increased to 181.7°C (Fig. 4). It should
be mentioned that the values of enthalpies of phase transition, ∆trH(γ→ α), can widely
differ for crystals prepared at slightly different conditions. For example, the γ-phase
having Ttr=165°C corresponds to a heat effect of 1200±80 J mol–1, whereas for
Ttr=180°C the heat effect was measured to be ∆trH(γ→ α)=1800±80 J mol–1.

Fig 4 DSC curves of the phase transition γ- →α-modification glycine for rhombus-like
crystals: 1 – the initial (fresh prepared) crystals, 2 – the crystals after thermal
treatment at 159°C for 5 h

Powder X-ray diffraction is usually used to identify the modification after the
phase transition [8]. In this study, identification of the transformed phase was carried
out in a different way. The same single crystal before and after DSC measurements
was studied by X-ray diffraction with the single crystal method. The results of these
investigations are presented in Tables 1, 2 and 3.
γ- → α-phase transition is irreversible and endothermic and can be characterised
as an enantiotropic phase transition, which is in accordance with the heat-of-
transition rule (HTR) and the infrared rule (IRR) introduced for polymorph transi-
tions by Burger and Ramberger [27]. The Gibbs’ free energies of the investigated
phases in the transition temperature are equal:
∆ tr H ( γ → α )=∆ tr G +Ttr ∆ tr S = 0−Ttr ( ∂∆G / ∂T ) p >0 (1)
in the point of phase transition Sα>Sγ. Assuming that volume and density are a mea-
sure of entropy, one can compare these values for different phases. The experimen-
tally obtained density of γ-form is higher than of α-modification:
ρ25(γ)=1.63 g cm–3 [8] >ρ25(α)=1.606 g cm–3 [28] or ρ25(α)=1.607 g cm–3 [29]. How-
ever, the values of the molecular volumes, Vmol, and the density calculated from X-ray
diffraction data, ρX-ray, just contradict the above relationship: Vmol(γ)=78.30(13)
Å3>Vmol(α)=77.92(17) Å3 and ρX-ray(γ)=1.592 g cm–3>ρX-ray(α)=1.600 g cm–3 [this

J. Therm. Anal. Cal., 66, 2001

 PERLOVICH et al.: THE POLYMORPHISM OF GLYCINE 709

work]. This fact may be explained by the essential influence of the network hydrogen
bonds in the overall organisation of the crystal lattices. Similar contradictions were
observed for another molecular crystal with network of hydrogen bonds [30].
As was mentioned before, the temperature of γ- → α-Gly phase transition is sen-
sitive to: a) form (geometry) crystals; b) thermal treatments; c) grinding and other
mechanical treatment. These facts essentially complicate kinetic and thermodynamic
investigations of this process. Taking this into account, in the present study it was at-
tempted to grow small crystals as homogeneously as possible with approximately the
same size, geometry and form. Moreover, it was tried to carry out all measurements
with the polymorphs under the same experimental conditions. For this goal, relatively
bigger crystals with different geometry (form) were grown: a) small rhombus; b)
small triangles; c) small pyramids; d) big pyramids, which were then carefully di-
vided into small pieces. The analysis of the kinetics of phase transitions for the differ-
ent crystal forms was performed as follows. The crystals were placed into a sealed
DSC pan and measurements carried out under isothermal condition as was used else-
where [31]. The Avrami–Erofeev equation [32] was used for the interpretation of the
time-dependency of the heat effect:
−ln(1−α * )= kt n (2)
where α* is the fraction of γ-phase transformed until time, t:
α * =∆ tr H (t )/∆ tr H sum (3)
∆trH(t) is the enthalpy of phase transition until time, t; ∆trHsum is the total enthalpy of
the phase transition. The index n in Eq. (2) was determined from the slope of the plot
of ln[–ln(1–α*)] vs. ln(t) according to the theory on nucleation-growth during crystal-
lisation [32].
It is interesting to note that for rhombus-like crystals the value n is Eq. (2), hence
it is observed a two-dimensional nucleus growth with a constant number of nuclei
(the nucleation rate is equal zero). For the triangle-like crystals, the value n increases
to 3. This corresponds to three-dimensional nucleus growth with a constant number
of nuclei (the nucleation rate is equal zero). Short periods of two-dimensional growth
and subsequent conversion into long period three-dimensional growth is observed
with pyramid-like small crystals. Similar behaviour is also observed for big pyra-
mid-like crystals: however, in some cases stages with n=3 substitute for n=4 (a con-
stant rate of homogenous nucleation and three-dimension growth). It is proposed, that
after a number of nuclei of the growing phase are finished, the mechanism of the reac-
tion of the phase transition is changed.
In the following part of this study, estimation of the activation energy of the phase
transition γ- → α-glycine is described. For this, isothermal measurements were carried
out at the different constant temperatures and dependencies of α*=f(T,t) were collected.
The activation energy at strictly fixed α* values, Ea(α*), was calculated from the slope of
the plot of ln(t α * ) vs. 1/T according to [32] (where t α * is a time of phase transition for

J. Therm. Anal. Cal., 66, 2001

710 PERLOVICH et al.: THE POLYMORPHISM OF GLYCINE

fraction reacted is equal α*). Dependencies between the ln(t α * ) vs. 1/T for the crystals
rhombus and triangle-like are listed in Figs 5 and 6, respectively.

Fig 5 Dependencies between the ln(tα * ) vs. 1/T for the crystals rhombus-like form:
*
n – α =70%; s – 75% ; o – 80%; o – 85%

Fig 6 Dependencies between the ln(tα * ) vs. 1/T for the crystals triangle-like form:
*
n – α =70%; s – 75% ; o – 80%; o – 85%

The results of the calculations of the activation energies of γ- → α-glycine phase

transition for the respective geometry of the crystals are shown in Table 4.

J. Therm. Anal. Cal., 66, 2001

 PERLOVICH et al.: THE POLYMORPHISM OF GLYCINE 711

Table 4 The activation energies, Ea, of the γ- →α-glycine phase transition for different geometry
of the crystals
Ea/kJ mol–1
α*/%
rhombus≡Ea(‘2’) triangle≡Ea(‘3’)
65 263±7 368±9
70 261±7 372±9
75 263±5 367±5
80 269±3 372±5
85 268±2 366±10
Ea 265±5 369±8

As can be seen from the table, the activation energy for three-dimension growth in
triangle-like crystals of α-phase, Ea(‘3’), is 1.4 times bigger than for the two-dimensional
nuclei growth in rhombus-like crystals, Ea(‘2’). If one takes into account Šesták’s [32]
approach, it may be proposed that the total activation energy of a phase transition process
is a sum of two terms: nucleation energy, EN, and growth energy, EG:
E a = E N +nE G (4)
where n is the order of the dimension of the growth process.
If a constant number of nuclei is assumed (EN → 0): Ea(‘3’)=3EG and Ea(‘2’)=2EG:
EG=(Ea(‘2’)+Ea(‘3’))/5≈218 kJ mol–1 (5)
Presumably the phase transition first takes place on defects crystal structure (Iitaka
[10] mentioned this fact). Therefore, the limiting stages of this process are the increase of
the defects and their diffusion to the reaction centres. Unfortunately, there are no publica-
tions estimating the energies of increasing and diffusion of different defects in organic
crystals, in particularly for glycine. However, it may be sufficient to roughly estimate the
kind of structure defects, which are responsible for γ- → α-phase transition: if the ener-
getic characteristics of nucleation and movement of the various defects in copper [33]
were used, the bivacancy mechanism would be more suitable.

The β-modification glycine

All the efforts in order to observe heat effects of the phase transition of the β-Gly in a
temperature interval of 25 up to 220°C (for different heating rate and isothermal treat-
ments) was not successful and no heat effects were observed. Moreover, a special ex-
periment was carried out as follows. The single crystal of β-modification was heated
from 25 up to 180°C (v =10 K min–1) and then isothermally treated for 30 min at this
temperature. After this procedure, the crystal was identified by X-ray diffraction. The
initial and tempered crystal had the same parameters of unit cell. According to Bur-
ger’s and Ramberger’s rules [27], the β-modification of glycine is monotropic with
respect to other phases.

J. Therm. Anal. Cal., 66, 2001

712 PERLOVICH et al.: THE POLYMORPHISM OF GLYCINE

It should be noted that in contrast to α- and γ-phases, the active sublimation pro-
cess was taking place in the temperature range from 210 to 240°C (v =10 K min–1)
with melting and decomposition at higher temperatures. Melting (and decomposi-
tion) temperature of α-glycine (v =10 K min–1) is 250°C. On the basis of these obser-
vations it may be assumed that the crystal lattice energy of β-glycine is less than that
of the α-modification. The confirmation of the fact will be shown in the next section
of the present study.

Solution calorimeter measurements

It should be mentioned that the investigation of isoenergetic polymorphs is a delicate
experimental problem. This is connected, first of all, with the fact that the difference
between the crystal lattice energies in many cases is of the same range as the experi-
mental errors. In this respect water was chosen as the solvent since the studied poly-
morphs dissolve in it better than in any other solvent. The mass of substance was ad-
justed such that the experimental signal was similar to the optimum signal for the
equipment used. The mass of each sample was close to 63 mg (m≈(8.4±0.1)·10–3) and
gave heat effects of about 12 J. The dilution enthalpy for concentrations between
ms=8.08·10–3 to mf=4.06·10–3 mol kg–1 is equal 0.8±0.4 J mol–1 [34]. The differences of
concentrations for the present experiments were less than 3% (m≈(8.4±0.1)·10–3).
Therefore, the dilution enthalpy correction to the solution enthalpy could be ne-
glected in the present study. In order to avoid mistakes due to artefacts under prepara-
tion of the crystals (size, defects, etc.), all the crystals were grown under the same
conditions and small single crystals of approximately identical sizes were chosen.
DSC and X-ray identified the investigated phases before the solution calorimeter
measurements.
The solution enthalpies of the examined phases are presented in Table 5.
Table 5 The solution enthalpies, ∆solHm, of the different polymorphic phases of glycine

α-phase β-phase γ-phase

N ∆solHm/ ∆solHm/ ∆solHm/
M/mg M/mg M/mg
J mol–1 J mol–1 J mol–1
1 63.468 14488 63.598 14170 61.856 14775
2 63.613 14481 63.794 14243 62.528 14925
3 63.347 14503 63.577 14180 62.998 14851
4 62.632 14502 63.390 14205 63.200 14795
5 61.527 14581 63.323 14190 63.049 14738
6 61.462 14561 63.857 14769
7 63.106 14523 63.121 14776
8 63.055 14636 64.074 14701
9 63.070 14433
∆ sol H m 14523±76 ∆ sol H m 14198±73 ∆ sol H m 14791±84

J. Therm. Anal. Cal., 66, 2001

 PERLOVICH et al.: THE POLYMORPHISM OF GLYCINE 713

It is well known that:

∆ sol H m =∆ sub H +∆ solv H m (6)
where ∆solHm is a solution enthalpy for concentration solute of m; ∆subH is a sublima-
tion enthalpy; Ecr is the crystal lattice energy (negative); ∆subH≈ –Ecr; ∆solvHm is a sol-
vation enthalpy for concentration solute of m.
∆ tr H ( A→ B )=∆∆ sol H m ( A→ B )=∆ sol H m (B )−∆ sol H m ( A )= (7)
∆ sub H (B )−∆ sub H ( A )=−∆E cr ( A→ B )
It should be emphasised that the differences between the crystal lattice energies
of the studied polymorphs may be small, but exceed the experimental error. In this re-
spect it should be mentioned that growth of the β-phase is in some cases observed
while preparing the α-phase. On the other hand, sometimes α-modification grows
when preparing γ-phase. However, the growth of a definite crystal lattice from a satu-
rated solution is presumably determined by the degree of association (the size and
structure of a cluster created in the solution). For example, the molecules of glycine
in an aqueous saturated solution exist mostly as dimers [35]. It may be assumed that
the structure of the crystal lattices of the various polymorphs reflects partly the struc-
ture of clusters in the saturated solutions. Based on analysis of the ∆solHm, the phases
can be arranged in a descending order of their absolute values of the lattice energies:
γ->α->β-modification.
As was elucidated above, the network of hydrogen bonds plays an essential role
in organisation of the crystal structure of glycine. Therefore, it is interesting to exam-
ine IR spectra of the polymorphic phases [7]. The absorption occurring at 3154 cm–1
in α-glycine corresponds to the highest N–H stretching frequency. This absorption
shifts to a lower value, 3093 in γ-glycine, and shifts to a higher value, 3180 cm–1, in
β-glycine. The shift in γ-Gly may be interpreted as an indication for the weakest hy-
drogen bond in the α-form being weaker than the weakest hydrogen bond in the
γ-form. The analogous situation reveals between β- and α-forms: the weakest hydro-
gen bond in the β-phase is weaker than the corresponding hydrogen bond in the

Fig 7 Dependence between the N–H stretching frequency and ∆solHm for the various
crystal modifications of glycine

J. Therm. Anal. Cal., 66, 2001

714 PERLOVICH et al.: THE POLYMORPHISM OF GLYCINE

α-phase. In β-glycine, each single layer is held to one another by weaker hydrogen
bonds than those contributing to the formation of double layers in the α-form. The
correlation between the N–H stretching frequency and ∆solHm examined on the three
phases of glycine is presented in Fig. 7. This result confirms once more the impor-
tance of the network of the hydrogen bonds for the crystal structure.

Conclusions
The temperature of the phase transition γ- → α-glycine turned out to be affected by a)
conditions under which the crystals of the γ-modification are grown, b) tempering of
the crystals prior to transition experiments, c) form (geometry) of crystals.
The kinetics of the phase transition γ- → α-glycine of rhomboedric single crys-
tals of γ-phase glycine is described best by the equation for two-dimension nuclei
growth, whereas for the crystals of triangle geometry it is the equation for
three-dimension growth.
Based on the energies measured for the growth of the new phase in the
γ- → α-phase transition, it may be supposed that a bivacancy mechanism of the move-
ment of crystalline defects would be plausible.
The network of hydrogen bonds plays an essential role in the organisation of the
crystal structure of glycine. Considering ∆solHm, the investigated polymorphs can be
arranged in descending order of absolute values of their lattice energies as follows:
γ->α->β-modification.
The β-modification of glycine is a monotropic phase with respect to the other
forms, whereas γ- and α-polymorphs are enantiotropically related.
* * *
This work was generously supported by Norges Forskningsråd, project number HS 58101.

References
1 J. D. Bernal, Z. Krist., 78 (1931) 363.
2 J. Hengstenberg and F. V. Lenel, Z. Krist., 77 (1931) 424.
3 G. Albrecht and R. B. Corey, J. Am. Chem. Soc., 61 (1939) 1087.
4 R. E. Marsh, Acta Cryst., 11 (1959) 654.
5 P. G. Jönsson and Å. Kvick, Acta Cryst., B28 (1972) 1827.
6 E. Fischer, Ber. Deut. Chem. Ges., 38 (1905) 2917.
7 Y. Iitaka, Acta Cryst., 13 (1960) 35.
8 Y. Iitaka, Proc. Jap. Acad., 30 (1954) 109.
9 Y. Iitaka, Acta Cryst., 11 (1958) 225.
10 Y. Iitaka, Acta Cryst., 14 (1961) 1.
11 Å. Kvick, Acta Cryst., B36 (1980) 115.
12 V. P. Vasil’ev, V. A. Borodin and S. B. Kopnyshev, Russ. J. Phys. Chem., (Engl. Transl.),
65 (1991) 29.
13 S. N. Ngauv, R. Sabbah and M. Laffitte, Thermochim. Acta, 20 (1977) 371.

J. Therm. Anal. Cal., 66, 2001

 PERLOVICH et al.: THE POLYMORPHISM OF GLYCINE 715

14 T. Tsuzuki, D. O. Harper and H. Hunt, J. Phys. Chem., 62 (1958) 1594.

15 V. G. Badelin, O. V. Kulikov, V. S. Batagin, E. Udzig, A. Zielenkiewicz, W. Zielenkiewicz
and G. A. Krestov, Thermochim. Acta, 169 (1990) 81.
16 C. H. Spink and I. Wadsö, J. Chem. Thermodynam., 7 (1975) 561.
17 J. O. Hutchens, A. G. Cole and J. W. Stout, J. Am. Chem. Soc., 82 (1960) 4813.
18 C. G. De Kruif, J. Voogd and J. C. A. Offringa, J. Chem. Thermodynam., 11 (1979) 651.
19 H. J. Svec and D. D. Clyde, J. Chem. Eng. Data, 10 (1965) 151.
20 S. Tagaki, H. Chihara and S. Seki, Bull. Chem. Soc. Jap., 32 (1959) 84.
21 J. D. Cox and G. Pilcher, Thermochemistry of organic and organometallic compounds, Aca-
demic Press, London 1970, p. 643.
22 G. M. Sheldrick (1997a). SHELXS-97 Program for the solution of crystal structures, Univer-
sity of Göttingen, Germany.
23 G. M. Sheldrick (1997b). SHELXS-97/2 Program for the refinement of crystal structures, Uni-
versity of Göttingen, Germany.
24 M. N. Burnett and C. K. Jonson (1996). ORTEP III: Oak ridge thermal ellipsoid plot program
for crystal structure illustration, Oak ridge national laboratory report ORNL-6895.
25 P. McArdle (1993). Oscail Software – Windows Software for crystallography from National
University of Ireland, Galway – Version 7d, J. Appl. Cryst., 26 (1993) 752.
26 K. Balasubramanian, R. S. Krishnan and Y. Iitaka, Bull. Chem. Soc. Jap., 35 (1962) 1303.
27 A. Burger and R. Ramberger, Microchim. Acta, II (1979) 259.
28 B. W. Low and F.M. Richards, J. Am. Chem. Soc., 74 (1952) 1660.
29 T. Curtius, J. Prakt. Chem., 26 (1882) 158.
30 A. Burger and R. Ramberger, Microchim. Acta, II (1979) 273.
31 G. L. Perlovich and A. Bauer-Brandl, J. Therm. Anal. Cal., 63 (2001) 653.
32 J. Šesták, Thermophysical properties of solids. Their measurements and theoretical thermal
analysis. Moscow, Mir 1987, p. 456.
33 C. E. Birchenhall, In: Reactivity of Solids, Elsevier, Amsterdam 1960, p. 24.
34 W. E. Wallace, W. F. Offutt and A. L. Robinson, J. Am. Chem. Soc., 65 (1943) 347.
35 R. M. Ginde and A. S. Myerson, J. Cryst. Growth, 116 (1992) 41.
36 J.-P. Legros and Å. Kvick, Acta Cryst., B36 (1980) 3052.

J. Therm. Anal. Cal., 66, 2001

View publication stats